Abstract
INTRODUCTION
The diagnosis of peri-prosthetic infection may be difficult. But this diagnosis can guide antibiotic prophylaxis and implementation of intraoperative bacteriological samples. The hypothesis of this study was that a composite score using clinical, radiological and biological data could be used for positive and negative diagnostic of infection before reoperation on prosthetic hip or knee.
MATERIAL
200 reoperations on hip and knee arthroplasty for any cause were analyzed retrospectively. 100 cases concerned infected cases, while the diagnosis of infection was excluded in the other 100 cases.
METHODS
Different criteria were collected: previous surgery of the operated joint, clinical, biological and radiological signs. Univariate analysis investigated the discriminating factors of infected and non infected cases. Multivariate analysis integrated concomitantly these factors. A composite score was defined, and its diagnostic efficacy was assessed by the percentage of the total variance explained by the score, by the percentage of correctly classified cases and by the sensitivity and specificity.
RESULTS
The univariate and multivariate analyzes have isolated the following significant factors: body mass index, diabetes, prosthetic mechanical complication, fever, existence of a wound defect. The composite score so defined allows separating the infected and non infected patients accurately in 78% of cases, with a sensitivity of 57% and a specificity of 93%.
DISCUSSION
The composite score defined predicts infection or no infection in the prosthetic joint before reoperation with good efficiency. This score could be a significant help to define the medical and surgical strategy in a prosthetic hip or knee reoperation for whatever reason.
CONCLUSION
A prospective study is needed to confirm definitively the contribution of this score.
